HKU Data Repository

File(s) under permanent embargo

Reason: All data in my thesis are or will be in the journal's review process.

Supporting Data for "Two Essays on the Adoption of New Technologies: How People perceive and Interact with Them"

posted on 2020-10-12, 01:14 authored by Xue Wang

In the recent years, different forms of new technologies, such as robots and artificial intelligence, have entered into our lives. Although behavioral scientists have started to investigate this phenomena, scholarly research in this field is still at an early stage. This thesis aims to advance our understanding on human interaction with technologies in business contexts.

Prior work has suggested the benefits of equipping robots with humanlike features in non-competitive contexts. Extending this, essay 1 examines robot anthropomorphism in a competitive context (i.e., workplace) and specifically focuses on the impact of humanlike features of robots on employees’ perception of job insecurity. Through two correlational and four casual experiments with samples from students, nurses, call-centre employees and crowdsourced workers, results show that equipping robots with humanlike features increases employees’ job insecurity perception due to an increasing social comparison with robots. These findings contribute to the literature on job insecurity by identifying a new antecedent of job insecurity, add to the ongoing conversation of robot anthropomorphism by examining it in a competitive context, and provide insights on the role of social comparison by applying it to human interaction with non-human agents. More importantly, implications are provided in this essay advising organizations on how to reduce employees’ job insecurity perceptions evoked by the adoption of robots from the design perspective.

Although previous literature has demonstrated what has led to algorithm aversion, very few research was found examining when it could be overcome. To address this research gap, essay 2 examines how exposure to moral violations impacts consumers’ preference for recommendation agents. Six experiments convergently demonstrate that exposure to moral violations enhances consumers’ preference for algorithm recommendation over human recommendation due to an increase of distrust in human and further an increase of trust in algorithm. Moreover, this effect is weakened when highlighting human agent’s trustworthiness or equipping algorithm with human intentions. These findings contribute to the literature on algorithm aversion byidentifying under which circumstances algorithm aversion could be reduced, extend previous literature on moral violations by investigating its influences on human interaction with non-human agents, and provide insights on the role of trust in recommendation adoption. In addition, these findings provide implications for marketers about when consumers’ algorithm aversion could be reduced.


Usage metrics

    Research Postgraduates


    Ref. manager